US child safety group NCMEC received 1.5M reports of suspected CSAM with ties to AI in 2025, a significant surge compared to 67,000 in 2024 and 4,700 in 2023
William Michael Haslach was a lunch monitor and traffic guard at a suburban Minnesota elementary school for years …LinkedIn:Kurt Wagner.
Context & Ripple Effects
Related coverage had already flagged that AI-generated abuse imagery could overwhelm the CyberTipline, while the Internet Watch Foundation documented a sharp rise in identified AI-generated abuse videos in 2025. Amazon’s reporting of potential CSAM found in AI training data adds a separate exposure point: harmful material can surface both in generation and in model-development pipelines.
The new NCMEC total makes the operational burden concrete. It also lands against a reporting system in which Meta’s very large volume of tips had already raised concerns among some prosecutors about investigative delays.
First-order effects
- NCMEC and downstream law-enforcement partners face a far larger AI-linked triage workload, increasing pressure to distinguish actionable reports from duplicative or lower-confidence signals.
- AI developers and platforms implicated in generation or training-data handling face more immediate scrutiny of detection, reporting, and provenance practices.
Second-order effects
- High report volumes can shift scarce investigative capacity toward prioritization and evidence validation, reinforcing the concern that reporting volume alone does not translate into faster case resolution.
- Platforms and model builders are likely to face stronger incentives to add safeguards across both user-facing generation systems and training-data ingestion, rather than treating moderation as a single downstream task.
Third-order effects
- If AI-linked reporting continues to grow, child-safety enforcement will increasingly depend on shared technical and procedural infrastructure for provenance, detection, and case prioritization—not only on expanding reporting channels.
- The pattern points toward public-safety governance becoming a core constraint on AI deployment, with scrutiny extending from model outputs to the data and operational systems behind them.
The trend: Generative AI is expanding the child-safety enforcement surface, forcing safety systems to adapt to harms that can be created, circulated, and detected at far greater scale.